Outlier Detection in Jeju International Airport Passenger and Flight Volumes - A Chen-Liu Approach

Abstract

This study aims to detect outliers caused by exogenous shocks in the monthly passenger volume and flight operations of Jeju International Airport and to identify their types. Monthly data covering 197 months from January 2010 to May 2026 were collected. A baseline SARIMA model was identified for each series, and the outlier detection procedure proposed by Chen & Liu (1993) was applied to classify outliers into additive outliers (AO), level shifts (LS), and temporary changes (TC). Seven outliers were detected in each series, and COVID-19 from 2020 to 2021 was confirmed as the dominant common shock. All seven outliers in the passenger series were downward and transitory in type, indicating a pattern of recovery following each shock. In contrast, the flight operation series exhibited both downward and upward level shifts, revealing an adjustment process in which operations declined sharply and then partially recovered, yet persisted at a level different from that before the shock. The 2015 MERS outbreak was detected only in the passenger series. These results demonstrate that an identical exogenous shock affects passenger demand and flight supply differently across three domains: timing, direction, and type. The findings suggest that aviation policy during crisis periods should be designed separately for the demand side and the supply side.